Persist info about subresources on pages for DNS pre-resolution

The DNS pre-resolution system already "learns" what domains are commonly
needed when rendering sub-resources of a page at a given domain.
This patch saves (some of) the information learned into a persistent
pref, and restores it on startup.

For now, I put in a wimpy pruning of the list each time I save, so that
the list will not grow endlessly from session to session.  I probably need
a better pruning algorithm, such as one that prunes after a given amount
of time, rather than only during shutdown.  For now, this should get
a lot of nice results, and provide slightly larger than needed lists to
users that have long lived sessions, which is similar to the current 
performance, where I didn't persist any info, and only pruned (actually
discarded) all learned info at shutdown.

r=mbelshe
Review URL: http://codereview.chromium.org/21374

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@9912 0039d316-1c4b-4281-b951-d872f2087c98

0 files changed, 0 insertions, 0 deletions